pping can crash if clock is adjusted backwards

[Gunnar Rangøy]

db.py:query:120 [Debug] Executing: SELECT rrd_fileid, netboxid FROM rrd_file
               WHERE path='/var/lib/nav/rrd' AND
                     filename='x.y.z.100.rrd'
Traceback (most recent call last):
  File "/usr/lib/nav/pping.py", line 271, in <module>
    start(nofork)
  File "/usr/lib/nav/pping.py", line 239, in start
    myPinger.main()
  File "/usr/lib/nav/pping.py", line 172, in main
    self.generateEvents()
  File "/usr/lib/nav/pping.py", line 117, in generateEvents
    rrd.update(netbox.netboxid, netbox.sysname, 'N', 'DOWN', 5)
  File "/usr/lib/pymodules/python2.5/nav/statemon/rrd.py", line 131, in update
    rrd.update(*rrdParam)
rrdtool.error: /var/lib/nav/rrd/x.y.z.100.rrd: illegal attempt to update using time 1306756336 when last update time is 1306757036 (minimum one second step)

[Morten Brekkevold (mbrekkevold)]

Christian, could you please look at this? I understand the exception thrown by rrdtool, this is normal behavior.

It is not acceptable for the pping daemon to crash because of this error. IMHO, it should log the incident and ignore it (meaning that no further updated will go into the RRD file until the clock passes the last update time again).